The Internet and Virtual Identities: A Corpus Analysis of Reader Responses to Interpersonal Violence
This paper examines a corpus constructed of online responses to an article in an online edition of a British tabloid newspaper describing an act of interpersonal street violence between two men. Taking a corpus-driven approach, the data was analysed by undertaking concordance analyses of keywords and collocates of those words. The findings indicate that, for certain individuals, interpersonal violence, and responses to acts of violence among men are tools for both the creation of and defense of self-image, and an acceptable avenue to accomplish masculinity. However, the findings also provide data revealing that other respondents reject such actions, clearly demonstrating that multiple constructs of masculine identity exist among the tabloid readership who responded to the article. The paper concludes by discussing the concept of online identity and online communities as well as the hypothesis that masculine identity and specifically hegemonic masculinity is constructed of multiple identities, and rejecting the notion that violence is a response to the destabilizing effects of post-modernism, while arguing that interpersonal violence is a means by which certain men express and validate masculinity.
Keywords: corpus linguistics; virtual identities; masculinities; interpersonal violence
The Internet and other Web-derived data have become a vast resource for corpus linguistics and natural language processing. The Web provides an unprecedented quantity of linguistic data from a broad range of registers and text types. In this study, texts of computer-mediated communication (CMC) taken from a message board of an online edition of a British tabloid newspaper, The Sun, are built into a corpus and analysed. The study researches the responses readers posted to an article in the newspaper which detailed an act of interpersonal street violence between two men in which one man was seriously injured and left unconscious in a street. This research utilises a corpus-driven approach in order to discuss the attitudes articulated by the posters towards the act of violence which allegedly took place, which it is argued, reflects upon the virtual identity of the posters.
As a result of the anonymity, freedoms of time and space, and the absence of audio-visual context on the Internet, virtual identity is considered to be more unstable, more performed and more fluid than ‘real’ identity, yet such a definition has similar qualities to postmodern identity which is described as both constructed and discursive (Bauman, 2007). Thus, an analysis of the interaction on an online message board focused on the topic of interpersonal street violence between two men may not only highlight the posters’ attitudes towards such violence, but furthermore, demonstrate traits of identity through discursive accomplishment.
The theme of the discussion board is centred on violence between two men. A great deal of what is bad in the world, from genocide to interpersonal violence, is the product of men and their masculinities (DeKeseredy and Schwartz, 2005). Work by criminologists such as Anderson (1990) have argued that instances of interpersonal violence originate from strongly held values in the construction and defence of personal street status and that violence is a tool for both the formation of and the protection of self-image. Furthermore, Messerschmidt (2004) writes that among certain men violence is a core component of masculinity and a means of proving one’s manhood. However, Winlow (2001) considers that street and pub fights function as a means for working-class men to actualise a masculine identity due to the loss of traditional industrial job opportunities in a postmodern society. Clearly, violence is one means by which certain men live up to the ideals of hegemonic masculinity; such practices may be learned through interactions with particular peer groups, or virtual peer groups.
The article which produced the data for this study was published in the online version of The Sun newspaper on January 8th, 2013. The Sun has the highest circulation in the U.K. and the tenth highest in the world with a daily circulation of over 2,400,000 copies. The newspaper is considered a tabloid paper as its format contains features such as sensational crime stories, gossip columns about the lives of celebrities and sports stars, and news stories many would consider sensationalist. Articles are often accompanied with large titles and photographs. The Sun’s target audience are people considered as working class and manual workers.
The article for this study, which was found in the News section, was titled, “Thug breaks man’s jaw outside takeaway in unprovoked attack…because he was ginger” below which were two pictures taken from CCTV footage, the first showing a larger man punching a second man. The second photograph shows the smaller individual falling to the floor in the street. After six sentences of the article, a CCTV video clip of the attack is embedded into the page for the readers to watch. Further down, there is another picture which depicts the larger man exiting a store and confronting the smaller man and a fourth photograph which shows the moment in which the smaller man was hit.
The article describes how a man was attacked and left seriously injured in what is described as an unprovoked attacked. The story contains a large proportion of direct quotes as the injured man describes the incident and the long period of physical and psychological recovery afterwards. The injured man described how he went into a pizza takeaway restaurant with his girlfriend and was sworn at and called a ginger pr**ck. He then states that he left the store but was attacked immediately outside. He was left unconscious with a badly broken jaw and needed three months to recover from the attack. In the article he is clearly depicted as the blameless victim, whereas the other man is presented as the guilty aggressor. The article states that the attacker was still being sought by the police at the time of publication.
The article produced 190 responses from readers containing 6,606 words. If a reader wished to comment on the message board, he or she would first have to create an account by either using an existing Twitter or Facebook account, or by creating a new account with the newspaper. The reader would have to submit a user name and if they wished an avatar. Using these two sets of information, 69 posters used male names or provided pictures of males, whereas only 4 posters indicated that they were female. The other poster provided user names and avatars which did not indicate gender.
The texts from the online message board were built into a corpus compatible with Wordsmith Tools software. In modern linguistics, a corpus is defined as a collection of authentic, computer-readable texts that are representative of a particular language or variety of language (McEnery et al. 2006: 5). According to Hunston (2002: 2), a corpus is not only defined by its form, but also by its purpose. A corpus is a planned collection of texts of naturally occurring examples of language which can be stored and retrieved electronically. It is designed for a linguistic purpose; this design determines the selection of texts used to compile the corpus.
Corpus linguistics should be considered as a methodology which can be employed in various areas of linguistic research rather than a linguistic theory (McEnery and Wilson 2001:1). The corpus linguistic analysis employed in this study is a corpus-driven investigation (Tognini-Bonelli 2001: 2). Utilising such an approach, the corpus is regarded as the data and the patterns within the corpus convey regularities in language, thus the analyst is committed to the integrity of the data as a whole (ibid: 84).
2.1 WordSmith Tools
WordSmith Tools (Version 6) is a suite of concordancing software developed by Mike Scott, distributed via the Internet. The package provides an integrated set of tools for analysing texts. In this study, the three principle tools of the software were utilised, they are as follows: The wordlist tool generates word lists in alphabetical and frequency order enabling texts to be compared lexically. It also provides statistics such as the total number of words. The keywords tool identifies words in a text whose frequency is unusually high in contrast with other texts. This allows a text to be characterised. The concordance tool generates a list of examples of where a particular word has been used in a text. It identifies collocates of the word and recognises common phrases.
Once the corpus of reader responses was compiled, the first stage of the analysis consisted of a study of frequency data. A frequency list is simply a list of all the types (words) in a corpus together with the number of occurrences of each type. Frequency is one of the most significant concepts which form the foundation for the examination of corpora. Frequency lists are essential tools for comprehending a corpus and are therefore a useful point to commence the analysis of any type of corpus. Wordsmith allows frequency lists to be compiled for single words and word clusters.
From the word list ordered by frequency, it was possible to gain an understanding of aspects of the corpus which occurred often and therefore had the potential to demonstrate the lexical choices that the tabloid readers who responded to the article made, which could relate to the presentation of particular discourses or attempts to construct identity.
Once a word list organised by frequency had been analysed, word lists arranged by keyness were observed. Words that are considerably more frequent in one corpus when compared against another (often a much larger reference corpus) are known as ‘keywords’. Wordsmith Tools contains a program which automatically carries out statistical tests on frequencies in two corpora, typically a smaller specialised one and a larger more general one. Comparing a smaller corpus to a larger reference corpus is a valuable means of formulating key concepts within the smaller corpus which makes it distinctive when evaluated against general language. For this study, a corpus of general English was constructed which consisted of newspaper articles from the British newspaper The Guardian.
When the corpora are compared, WordSmith lists the keywords for the specialised corpus. The keyword procedure can therefore be used to identify the significantly different lexis between the specialised corpus and the larger general corpus. A keyword list is to be expected to be more useful in signifying lexical items that possibly will merit additional investigation than a raw frequency list as a keyword list provides a degree of prominence, instead of frequency alone. WordSmith calculates the size of each corpus and the frequencies of the words within them. In this study, the program was set to perform a log likelihood statistical test for each word, which gave a probability value (p value). This value designates the degree of confidence that a word is key due to chance alone, the smaller the p value, the more likely that the word’s presence in one of the corpora is not due to chance but the result of the author’s choice to use the word consciously or subconsciously. Wordsmith uses a default of p<0.000001.
Once the keywords lists had been studied, to gain a more comprehensive insight into the data, the keywords had to be observed in context by undertaking a concordance analysis. A concordance analysis combines quantitative and qualitative analysis and therefore may be considered as more productive than relying on quantitative analysis alone. Baker (2006: 71) writes, “A concordance analysis is one of the most effective techniques which allow researchers to carry out this sort of close examination.” A concordance is a list of all the occurrences of a particular search term in a corpus presented with words to their left and right as they occur, thus in context. The WordSmith Tools package contains a concordancer which places the selected word, known as the node word, in the centre of the screen with words that come before and after it on the left and right respectively.
A concordance-based study is able to disclose a range of discourses; therefore the notions of semantic preference and semantic prosody are important concepts. Semantic preference is defined by Stubbs (2001: 65) as “the relation between a lemma or word form and a set of semantically related words”, thus it is related to the notion of collocation. Semantic preference is the meaning which arises from the common semantic features of collocates of a given node word (McEnery et al. 2006: 84). For example, in the COCA Corpus, the word man co-occurs with words to do with appearance: e.g. distinguished-looking, big-bellied, grey-haired, ruddy-faced, heavyset, bearded etc. It does not necessarily collocate with just a single word, but with various lexical sets of semantic categories. Semantic preference is linked to the notion of semantic prosody (Louw, 1993) where patterns in discourse can be established between a word and a set of related words that indicate a discourse.
To gain further insights into the usage in context of the keywords, a study of collocates of those words was undertaken. Analysing collocation is a means of discerning meanings and associations between words which are otherwise difficult to ascertain from a small-scale analysis of a single text. The context in which words are situated is of paramount importance when the construction of meaning is considered. Words can be found in characteristic collocations, such collocations demonstrate the associations, and connotations they encompass, and as a result the suppositions that they represent. Thus, the analysis of collocates facilitates the semantic comprehension of a word. It contributes to an understanding of both semantic preference and semantic prosody.
A technique known as mutual information (MI) was utilised to calculate the strength of collocation. This method takes into account the extent to which two words occur apart from each other as well as together. Thus, the MI score measures the degree of non-randomness present when two words co-occur. An MI score of 3 or higher can be considered to be significant (Hunston 2002: 71). However, one problematic issue with MI is that high scores may be achieved by relatively low frequency words, therefore this must be taken into consideration as words with a low frequency ought not to be considered as significant in spite of a high MI score.
The following section describes the findings of the analysis using the methodologies described above.
Frequency is one of the most fundamental concepts in the analysis of a corpus which is able to provide insights which illuminate a range of themes. Table 1 presents the twenty most frequent words in the corpus.
Table 1. Twenty most frequent words
It is apparent that the most frequent words in the corpus are grammatical words (function words). Such words belong to a closed grammatical class consisting of high frequency words such as articles, pronouns, conjunctions and prepositions. These groups of words do not necessarily provide insight to the discourses found within the corpus as most forms of language contain a high proportion of functional words. However, by taking into consideration the most frequent lexical words such as nouns, verbs, adjectives and lexical adverbs, a clearer notion of the discourses within the corpus is attained.
Table 2. Twenty most frequent lexical words
Table 2 presents a much clearer picture of what the corpus is about. There are words describing an act of violence (victim, punch, fight). Another group of words are used to describe a person negatively (thug, scum, coward). An additional aspect of the list is that catch and caught are both present, therefore the lemma CATCH is significant on the frequency list and thus within the corpus. When this is taken into consideration with prison, it can be seen that another theme is prominent in the corpus. Beside the lemma CATCH and punch, other verbs are also present (like, get, hope, know, got, think). Further analysis of collocational data and concordance lines will be needed to understand the context of these verbs within the corpus, although when the COCA corpus of general English is referenced, it can be seen that such verbs are also of high frequency in a corpus of general English.
It is also possible to consider frequency lists beyond the single word. Table 3 presents the most frequent 3-word clusters.
|1||is going to||7|
|2||he is a||6|
|3||lock him up||6|
|4||him up and||5|
|5||on the wrist||5|
Table 3. Most frequent 3-word clusters
Although it is not possible due to space restrictions to present the context for each of the word clusters, by observing the concordance lines for the most frequent cluster is going to, certain themes can be seen.
1 suspended, you just know He is going to get off lightly
2 moment of madness and this kid is going to pay the ultimate price
3 punching someone. This place is going to the dogs at a rapid rate
4 the lad who threw the punch is going to have to deal with the ‘victim’s friends
5 going to be mugged or someone is going to attack me, so im always ready
6 kick someones head in . Apparently UKIP is going to sort this out….!
7 just a stupid no brain thug who is going to jail. I dont think
The writer of line 1 predicts the assailant is not going to be severely punished, whereas in lines 2 and 7 the opposite prediction is made. Line 4 remains within the theme of punishment, but states that the attacker is going to have to face the victim’s friends. In line 3, the poster describes the decline of society and social behaviour, which is similar to the sentiment found in line 6, which states that a British political party claims to have a solution for such a situation, although the use of the word apparently appears to contest such a claim. The poster of line 5 describes the actions that he or she would take if attacked in similar circumstances, thus claiming to be more prepared to act in the instance of street violence than the victim was. Such examples highlight the diverse responses to the act of violence depicted in the article.
Of the six instances of he is a, five refer to the attacker. He is described as: a self centred thug, a threat to the society and a trained fighter. Although further analysis is necessary, the data demonstrates certain themes within the corpus; the assailant is condemned for his action and his fighting ability is discussed. Two clusters are evident which combine to produce the phrase: lock him up and followed by a phrase such as throw away the key demonstrate the punishment the posters consider the aggressor deserves. Another frequent cluster is part of the phrase: (slap or smack) on the wrist, thereby predicting that he will be dealt with lightly by the law. Thus it can again be seen, there are various reactions and opinions to the violence: the aggressor is condemned for his actions, that society is described as having poor moral standards, and that he will not receive adequate punishment for his actions.
Therefore, it can be observed that by analysing frequency lists, discourses within the corpus may be highlighted. As has been described, the following discourses have emerged: the aggressor is condemned for his actions, society is seen as being in moral decline, the law is inadequate, and the fighting ability of the attacker is commented upon.
In the following section, the keywords of the corpus will be discussed.
Table 4. Keywords ordered by keyness
Table 4 presents the keywords with the highest levels of keyness. These words could be divided into three separate groups. There are functional words: him, this, I, he, why, your, don’t, out, verbs: hope, get, like, catch, and nouns or adjectives: guy, someone, thug, victim, punch, ginger, scum, coward. It is not possible to present all the findings in this paper due to space restrictions, therefore certain keywords from each will group be selected for further analysis.
If the functional words are taken into consideration, him has the highest level of keyness and the second most frequent among the keywords on the list. As this word is also most likely to be referring to one of the two male actors in this instance of street violence, further analysis may provide further insights as to how the two men are constructed. Therefore, the keyword him was studied in context by considering the concordance lines.
Of the 84 instances of him, 73 are referring to the attacker, of which 36 depict him negatively, 10 reference the victim and one refers to a person in a hypothetical situation, thus the aggressor appears to be the primary focus of the posters. If the collocates of him with the strongest levels of MI scores are calculated, the following list is provided: suspended, teach, catch, example, sentence, years, really, throw, someone, prison. This appears to indicate that within the corpus there is a dominant discourse associated with the attacker being caught and punished for his actions. This is confirmed when the word him is seen in context. A principle discourse focuses on the attacker being caught: Catch him and jail him ASAP. Another concordance line within the same semantic field describes the same notion more strongly: Sum of Britain!!!PLEASE catch him. Another example using a phrase which was found in the most frequent clusters is as follows: Find that punk, lock him up and throw away the key. Therefore it can be seen how the posters react to such acts of violence. Another discourse within the corpus denigrates the attacker, as the last example demonstrates with the term punk. Other examples include: UK is full of scum like him! / No other word for him COWARD., and referring to him as a mug brained idiot. However, not all of the posters refer to the aggressor in such negative terms, nor do all the people who responded to the article believe that he should receive a prison sentence for his actions as the following examples indict: but the other kid does square up to him / I doubt he wudda smacked him like that completely unprovoked / What if the ‘victim’ offered him out to begin with? This appears to indicate that there are some writers who do not accept the opinion of the article and are willing to consider alternative scenarios for the event which took place and this indicates that the response to acts of violence among the message board posters is not uniform or homogenous.
Another keyword which provides insights into the corpus is punch. There are 19 instances of this word, all of which are in the form of a noun. When analysed in context, opposing discourses are evident; 9 of the lines either defend the attacker or are appreciative of his fighting skills, whereas only 5 instances denounce his actions. Another 5 instances of punch were classified as neutral, neither defending nor denouncing the attack. Examples of instances which depict the attack positively are as follows: Boom! What a punch! / great punch, and the kid knows how to throw a punch. Such examples are in contrast to the article which clearly denounced his actions. It can be seen that the writers of these examples value the act of violence regardless of the fact that it left one man seriously injured. As previously stated, the corpus does not contain a single discourse; other examples of punch in context are more condemning: The guy who threw the punch is a bully / it was a dangerous cowardly sucker punch. This brief study of punch demonstrates that both qualitative and quantitative analysis is necessary not only to discover discourses within a corpus, but also to comprehend their statistical significance.
Another keyword of interest is victim. There are 25 instances of this word; they all refer to the man who was left unconscious with a broken jaw. However, when the concordance lines are studied, it can be seen that a number of writers are using this word ironically when labelling this man as a victim of crime or cast doubt on the interpretation of events depicted by the newspaper. Seven of the writers do not consider the injured man to be blameless as the following examples demonstrate: the person who hit the deck was not a victim / Looks like the ‘victim’ called the other guy out of the joint then got punched / I doubt that he’s the complete victim he’s made himself out to be. These writers do not appear to accept the opinions of the newspaper nor the evidence provided by the link to the CCTV footage which clearly illustrates the assault. The writer of the second example writes the word within quotation marks to emphasis the fact that it is doubted whether the person is in fact blameless. Other writers demonstrate a different opinion, as the following examples illustrate: it looks like he had 20lbs over the victim / That victim could have been a brain op patient.
The data demonstrates that the writers who posted on the newspaper message board in response to the article are clearly of differing opinions. There are those who accept the views presented by the newspaper which condemn the assailant and his actions, clearly articulating how he should be punished as a consequence of his actions. Others use the incident to express an opinion the England has and is still experiencing a decline in social standards and morality, and furthermore that law enforcement is too lenient to effectively respond to such a situation. However, there is another statistically significant discourse within the corpus which is contrary to those which condemn or criticise the violence. In this discourse, violence is seen as something which is appreciated and respected, where the aggressor is not depicted as the guilty party and where the victim is not seen as blameless. It is of interest that such opinions of violence are articulated on a message board which is overwhelmingly populated by writers who claim to be male, thus the data provides insights into how certain individuals construct and respond to identities of masculinity.
The findings have demonstrated a continuum of opinions and stances on the online message board in response to a specific act of violence. Such expressions may be considered to be a reflection of an aspect of identity the writer constructs for himself or herself. Early CMC scholars described how the Internet liberated people from social constrain through there being a supposedly unbiased and non-prejudiced environment. The Internet was also believed to provide a measure of anonymity; however this perception has now appeared to have lessened due to the rise of social media networks in which any form of Internet activity is traceable and where users are aware of a degree of accountability regardless of the spatial distance when interacting on the Internet (Thurlow et al 2004).
As the majority of posters on the message board claimed to be male, masculinity is an important concept in this study which can be defined as the trait of behaving in ways that society considers to be typical and acceptable for males. Masculinity, like gender, is constructed and therefore is something that has to be worked at. Boys and men have to prove their masculinity constantly (Kimmel 2001: 269). Masculinity is often defined in terms of what it is not. Therefore it is often set up in relation to other identity components, particularly femininity.
Hegemonic masculinities (Connell 1995) are characterised as the variety of masculinity capable of marginalising and dominating not only women, but also other men. It is dependent on subordinate masculinities, since it must contradict them. These subordinate masculinities need not be clearly defined; in fact, it may be advantageous for hegemonic masculinity that they are not. However, in the data presented in this paper, in can be seen that subordination is achieved through violence, although this form of action is rejected by certain men. For certain researchers, such as Whitehead (2002: 93-94), discourse is focused upon as a means to comprehend how men practice hegemonic masculinity and perform identity work.
Masculine identities can therefore be understood as effects of discursive practices; they are fashioned within institutions and are historically constituted. One way that the gender order is maintained is by linking notions of appropriate and inappropriate gendered performances to different types of identities. The notion that masculinity is a singular rather than multiple identities has been viewed as problematic, particularly where gender identities and power relations are contextualised practices. Whitehead (2002: 33-34) writes:
It is no longer tenable, given recognition of the multiplicity, historicity and dynamism of gender representations, to talk of masculinity in the singular. Rather, we can see that masculinities are plural and multiple; they differ over space, time and context, are rooted only in the cultural and social moment, and are, thus, inevitably entwined with the powerful and influential variables such as sexuality, class, age and ethnicity.
In order to comprehend the diversity of masculinities, it is necessary to study the relations, such as subordination and dominance, between the different forms of masculinity. These relationships are constructed through practices that may intimate or exploit others (Kiesling 2006: 118). Masculinity is not a fixed trait, but a social process dependent upon restatement, and which, in various forms, involves language, thereby centrally situating linguistic issues in the theorising of gender. Men who heavily invest in a particular masculinity will attempt to communicate in a manner particular for that specific trait (Moita-Lopes 2006: 294). Masculinities are not displaced from a social context, but embedded and implicated in the lives of men.
A person’s identity is affected by the social interaction made and the stances and opinions expressed both in an offline and online environment. It could be argued that Internet culture has progressed to an extent that online identities reflect offline identities due to the diminished time and space restrictions which allow an individual to locate a group or topic online with which to associate and interact with, consequently allowing the individual to explore and construct an identity with that discussion forum or group. The Internet does not alter the approach by which identity is created; it merely provides another means through which identity construction can occur.
Responses to acts of violence may be considered by certain individuals to be a tool for both the creation of and defence of self-image. Using corpus linguistic methods, the data has highlighted discourses which demonstrate that a wide range of stances exist, which in turn signifies the plurality of masculinity, assuming that the message board analysed in this study was overwhelmingly populated by men. Social action is the framework of social structure; it is in the course of the performance of gendered behaviour that gendered social structure is reproduced. This entails that gender is relevant to all actions and activities. Among the texts analysed in the data, it has been shown that there exists a discourse, and therefore a subculture of violence, whereby acts of aggression are respected and esteemed, and therefore a means to construct a masculine identity.
Researchers such as Messerschmidt (2004) argue that with the loss of traditional industrial job opportunities and the shift towards a service-based economy, certain working class men have found new means of establishing masculinity; violence and street fights are one means of doing so. Such a form of masculinity emphasises toughness and a willingness to fight and defend oneself in the face of perceived threats or challenges by other males. The findings have shown that all men will not respond to violence in the same way. This will reflect upon an understanding of what men are and the consequences of acts such as the one focused upon in this study. However, this study has shown that for certain men, violence is considered a technique for validifying masculinity through peer support which encourages and legitimises acts of aggression. Hegemonic masculine discourses and practices, such as violence, may be learned through interactions, both virtual and face to face, which justify the relevance of studying online communication.
Although researchers such as Winlow (2001) consider violence amongst males to be a consequence of the destabilising effects of postmodernism, others such have Pinker (2011) describe how violence has been a constant trait throughout human history. It may be argued therefore that violence and aggression by men is more closely linked to aspects of patriarchal and hegemonic masculinity (Connell 1995) than it does with social responses to the effects of postmodernism.
The Internet is a location where individuals may construct their identity by expressing stances and through interactions with other Internet users. Whenever an individual interacts in a social environment, an aspect of their identity is revealed, and as identity construction and maintenance is a continual process, further construction takes place; this also applies for an online environment. The identities that individuals construct and the interactions they make on locations such as message boards may not necessarily be totally reliable or accurate. However, Wiszniewski and Coyne (2002) argue that regardless of the reliability of the interaction or identity construction, a reflection of the authentic identity is formed which will reveal an aspect of the user’s identity.
This paper has demonstrated that by employing a corpus linguistic approach, multiple expressions of identity and identity construction on the Internet may be studied. Discourses of violence and masculinities have been discussed, and the continuum of responses to violence observed and analysed. The results indicate that for certain individuals violence and aggression is an esteemed character trait, while for others it is rejected and condemned, thus confirming the notion of multiple masculine identity traits rather than a singular stereotypical construction.
Baker, P. (2006). Using Corpora in Discourse Analysis. London: Continuum.
Bauman, Z. (2007). Liquid Times. Cambridge: Polity.
Connell, R. W. (1995). Masculinities. Cambridge, Polity.
DeKeseredy, W.S. & Schwartz, M.D. (2005) Masculinities and Interpersonal Violence. In M.S. Kimmel, J. Hearn, R.W. Connell (Eds.) Handbook of Studies on Men & Masculinities: SAGE Publications: London. (pp. 353-366).
Firth, J.R. (1957). Papers in Linguistics 1934-1951. London: University Press.
Hunston, S. (2002). Corpora in Applied Linguistics. Cambridge: Cambridge University Press.
Kiesling, S. F. (2006). Playing The Straight Man. In D. Cameron & D. Kulick (eds.) The Language and Sexuality Reader. Oxford: Routledge. pp. 118-131.
Kimmel, M. S. (2001). Masculinity as Homophobia: Fear, Shame and Silence in the construction of Gender Identity. In S. M. Whitehead & F. J. Barrett (eds.) The Masculinities Reader. Cambridge: Polity. pp. 266-287.
Louw, B. (1993). Irony in the text or insincerity in the writer? The diagnostic potential of semantic prosodies. In Baker, M., Francis, G. & Tognini-Bonelli, E. (eds.) Text and Technology: In honour of John Sinclair. Philadelphia and Amsterdam: John Benjamins, 157-176.
McEnery, T. and Wilson, A. (2001). Corpus Linguistics. An Introduction. Edinburgh: Edinburgh University Press.
McEnery, T., Xiao, R. and Tono, Y. (2006). Corpus-Based Language Studies: An advanced resource book. London and New York: Routledge. Messerschmidt, J. (2004) Flesh and Blood: Adolescent Gender Diversity and Violence: Lanham: Rowan and Little Field.
Moita-Lopes, L. P. (2006). On being white, heterosexual and male in a Brazilian school: multiple positionings in oral narratives. In A. de Fina, D. Schiffrin & M. Bamberg (eds.) Discourse and Identity. Cambridge: Cambridge University Press.
Pinker, S. (2011). The Better Angels of Our Nature: Why Violence Has Declined. New York: Viking.
Stubbs, M. (2001). Words and Phrases. Oxford: Blackwell.
Thurlow, C., Lengel, L., & Tomic, A. (2004). Computer Mediated Communication: Social Interaction and the Internet. Thousand Oaks, CA: Sage Publications.
Tognini-Bonelli, E. (2001). Corpus Linguistics at Work (Studies in Corpus Linguistics: 6). Amsterdam/Atlanta, GA: John Benjamins.
Whitehead, S. M. (2002). Men and Masculinities : key themes and new directions. Cambridge: Polity Press.
Wiszniewski, Dorian. , & Richard Coyne (2002), Mask and Identity: The Hermeneutics of Self-Construction in the Information Age. In K. Ann Renninger & Wesley Shumar (Ed.) Building Virtual Communities (pp. 191-214). New York, New York: Cambridge Press.
Winlow, S. (2001) Badfellas: Crime, Tradition and New Masculinities: Berg: Oxford.
 “ABCs: UK National Newspaper Sales 2013”. Press Gazette (UK).
 “The Sun – Facts and Figures”. Newspaper Marketing Agency
 The frequent appearance of a word either near or next to another word creates a relationship between the two words which is labelled as collocation (Firth 1957).
 The data is present in the form in which it was collected.